Flooding Behavior Near The US/Canada Border: Investigating Forecasting Complications And Potential Resolutions

Academic Level at Time of Presentation

Senior

Major

Computer Science

Minor

Data Analytics

List all Project Mentors & Advisor(s)

Dr. Jeff Osborne; Dr. Zheng Zhang

Presentation Format

Oral Presentation

Abstract/Description

Flood forecasting remains a major challenge due to the nonlinear nature of hydrological systems and uncertainties in environmental data. This study aimed to address the prevalent challenges that arise from forecasting flooding behavior. To address the inherent complexity of hydrological forecasting, a machine learning framework was developed and trained on major contributing factors. To achieve an optimal balance between computational efficiency and predictive performance, a Gated Recurrent Unit (GRU) was selected as the optimal machine learning model. As the chosen dataset, North American Land Data Assimilation System Phase 2 (NLDAS2), is known to have inaccuracies in the important feature Relative Soil Moisture (RSM) around the USA borders, a river gauge affected by this inaccuracy was selected. Cumulative Distribution Function (CDF) matching was utilized before training with the aid of the unaffected Global Data Assimilation System (GDAS) dataset. Then, the preserved transformation function was saved and applied before RSM was input into the model. Finally, the model’s performance was analyzed to assess overall predictive success and the impact of CDF matching on model accuracy.

Fall Scholars Week 2025

Honors College Senior Thesis Presentations

This document is currently not available here.

Share

COinS
 

Flooding Behavior Near The US/Canada Border: Investigating Forecasting Complications And Potential Resolutions

Flood forecasting remains a major challenge due to the nonlinear nature of hydrological systems and uncertainties in environmental data. This study aimed to address the prevalent challenges that arise from forecasting flooding behavior. To address the inherent complexity of hydrological forecasting, a machine learning framework was developed and trained on major contributing factors. To achieve an optimal balance between computational efficiency and predictive performance, a Gated Recurrent Unit (GRU) was selected as the optimal machine learning model. As the chosen dataset, North American Land Data Assimilation System Phase 2 (NLDAS2), is known to have inaccuracies in the important feature Relative Soil Moisture (RSM) around the USA borders, a river gauge affected by this inaccuracy was selected. Cumulative Distribution Function (CDF) matching was utilized before training with the aid of the unaffected Global Data Assimilation System (GDAS) dataset. Then, the preserved transformation function was saved and applied before RSM was input into the model. Finally, the model’s performance was analyzed to assess overall predictive success and the impact of CDF matching on model accuracy.